On Thu, Oct 06, 2005 at 11:42:50AM -0400, Sam Carmichael wrote: > Hi, > > I'm new to ganglia. I'm wondering if anybody use it to do monitoring > on a per-process basis than per-machine basis. Suppose I have a > cluster of machines and each machine runs multiple instances of a > particular process, and I want to publish gmetric stats from each > process. From what I can see, the gmetric structure is rather flat, I > don't quite see how to create a hierarchy of metrics that goes from > machine->process->metric. I can prepend the process instance name > to the metric, but this inevitably increase the metric message size > rather significantly (I have many metrics per process).
There is no hierarchy to the metrics. A hierarchy would be a useful feature. -- Brooks -- Any statement of the form "X is the one, true Y" is FALSE. PGP fingerprint 655D 519C 26A7 82E7 2529 9BF0 5D8E 8BE9 F238 1AD4
pgpbKRBpCR46k.pgp
Description: PGP signature
